Positive Direct Coombs Test Induced by Phenylhydrazine

E. E. Muirhead, … , M. Groves, S. Bryan

J Clin Invest. 1954;33(12):1700-1711. https://doi.org/10.1172/JCI103050.

Research Article

Find the latest version: https://jci.me/103050/pdf POSITIVE DIRECT COOMBS TEST INDUCED BY PHENYLHYDRAZINE 1 By E. E. MUIRHEAD, M. GROVES, AND S. BRYAN (Froim the Department of Pathology, Southwestern Medical School of the University of Texas. Dallas, Tex.) (Submitted for publication April 9, 1954; accepted July 21, 1954)

The Coombs or antiglobulin test was introduced in 95 per cent ethyl alcohol and the percentage concen- for the detection of antibodies to erythrocytes of tration of erythrocytes containing these structures was a type termed "incomplete antibodies" (1). The obtained by counting 1000 erythrocytes (27). The me- chanical fragility of erythrocytes was determined by a indirect Coombs test has gained wide application modification of the method of Shen, Castle, and Fleming in the detection of the antibodies to various eryth- (28). One cc. of oxalated blood (5 cc. blood plus 10 rocytic antigens (2) and in the cross-matching of mg. of a buffered mixture of potassium oxalate and am- blood for transfusions (3). The direct Coombs monium oxalate) was placed in a 25 cc. Erlenmeyer flask test has become a diagnostic and investigative tool containing 20 glass beads, each of about 4 mm. diameter. The flask was clamped to a plastic disc 16.5 cm. from in hemolytic disease of the newborn (4-6), fol- the axis and rotated for 90 minutes at 30 rpm. at room lowing incompatible blood transfusions (7) and temperature. The amount of hemoglobin in the superna- in the acquired hemolytic anemias, idiopathic or tant plasma was related to the total hemoglobin concen- secondary types (8-22). tration of the blood as per cent hemolysis. All of these conditions in man when associated The plasma hemoglobin concentration was determined by the method of Bing and Baker (29). The serum bili- with positive direct and indirect Coombs tests, rubin concentration was determined by the method of have been interpreted to be the result of immune Malloy and Evelyn (30). mechanisms, either of the isoimmune or autoim- The Coombs tests: The canine serum anti-serum was mune type. In the present communication the pro- prepared in the rabbit by slight modifications of the di- duction of a positive direct Coombs test in the rections of Mourant (31): Rabbits weighing 2 to 3 kg. were injected with pooled canine serum obtained by mix- dog by means of the drug phenylhydrazine is de- ing about 5 cc. of serum from each of three to five dogs. scribed. The phenomenon is unattended by the A course of intravenous injections was given at 2-to-3- immunologic implications of blood incompatibility day intervals beginning with a 0.5 cc. dose and following and is not associated with the usual conditions with 5 doses of 1 cc. each. The animal was bled 10 a or days after the last dose. Pooled erythrocytes from nor- causing secondary symptomatic acquired he- mal dogs were used in absorbing the rabbit serum. The molytic anemia. serum was then passed through a Seitz filter and ali- quots of 1 cc. were stored in small sterile bottles at MATERIAL AND METHODS -20'C. until used. The results with this serum were on nine occasions with the results a Adult normal mongrel dogs of either sex weighing compared using canine serum furnished Drs. L. E. 5.5 to 11 kg. were used. After the control observations antiglobulin kindly by had been completed, a 1 to 2 per cent solution of phenyl- Young and S. N. Swisher of Rochester, New York. hydrazine hydrochloride (Merck) in saline was injected When these two sera were used undiluted with cells giv- intravenously in a single dose of 40 mg. per kg. body ing positive results, the degree of agglutination evoked by weight. The studies were conducted periodically there- the two was identical. after up to 41 days. In the direct Coombs test one drop of a 2 per cent The hemoglobin concentration of the peripheral blood suspension of the canine erythrocytes was washed three was determined by the alkaline hematin method (23). times in abundant saline at room temperature. After the The hematocrit was determined according to Wintrobe third wash the test tube was turned up and allowed to (24) and the reading was corrected for trapped plasma drain. To the erythrocytes suspended in the residual according to the method of Chaplin and Mollison (25). saline, a suspension approximating the original 2 per cent Reticulocytes were estimated routinely by counting 1000 concentration, 2 drops of the canine serum anti-serum erythrocytes stained with brilliant cresyl blue (26). (Coombs serum) were added. The mixture was al- Heinz bodies were stained with 0.2 per cent methyl violet lowed to stand for 30 minutes following which it was centrifuged for one minute at 500 to 1000 rpm. After 1 Supported by a grant from the U. S. Public Health shaking the button of cells into the supernatant the re- Service, National Heart Institute. sults were evaluated grossly as follows: a solid button 1700 POSITIVE DIRECT COOMBS TEST INDUCED BY PHENYLHYDRAZINE 1701 of agglutinated cells was designated 4 +, a few large The indirect Coombs test was conducted by incubating clumps 3 +, multiple scattered smaller clumps 2 +, a erythrocytes from normal dogs as a 2 per cent suspen- definitely granular suspension with fine clumps 1+. All sion at 370 C. for 30 minutes in the serum of the dogs of these grossly detectable degrees of agglutination were receiving phenylhydrazine and then treating the cells as checked by microscopic inspection which also demon- in the direct procedure. A random panel of canine strated that the clumps of erythrocytes were stable for erythrocytes obtained from three normal dogs was used periods longer than 5 minutes. A smooth or fairly throughout any one experiment. smooth suspension which upon microscopic inspection The titration of the Coombs serum, with minimal mod- yielded definite but scattered clumps was designated +. ification, was conducted in accordance with the suggestion Rare microscopic clumps in a grossly negative prepara- of Evans and Duane (11). Serial dilutions of the tion were ignored and the test was called negative. Since canine serum anti-serum with saline were prepared. To the + designation is questionable, no significance is given 0.1 cc. of the diluted serum was added 0.05 cc. of a 2 to it in this presentation. The microscopic appearances per cent suspension of washed erythrocytes. The mix- are depicted in Figure 1. ture stood at room temperature for 30 minutes, then was

FIG. 1. MICROSCOPIC APPEARANCE OF THE VARIOUS GRADES OF AGGLUTINATION INDUCED BY THE DI- RECT COOMBS TEST FOLLOWING PHENYLHYDRAZINE. (A) DEPICTS THE MICROSCOPIC APPEARANCE OF A GROSS ± REACTION; (B) A 1 + REACTION; (C) A 2 + REACTION; AND (D) A 3 + REACTION 1702 E. E. MUIRHEAD, M. GROVES, AND S. BRYAN centrifuged at 500 to 1000 rpm. for one minute and evalu- above. The hemolyzed erythrocytes were prepared as ated for clumping. follows: the erythrocytes from the normal animals used The trypsin-treated erythrocytes were prepared by a were washed three times with abundant saline, a final sus- modification (32) of the description of Morton and Pickle pension of erythrocytes in saline was prepared with the (33). A 2 per cent solution of trypsin (Difco 1: 250) in same concentration of erythrocytes as originally present buffered saline (pH 7.3) was stcred at - 200 C. in small in the blood, this suspension was hemolyzed by repeated aliquots until used. A 1: 10 dilution of this 2 per freezing and thawing. Altogether there were eight tests cent solution in buffered saline was used as a working against the sera and eight tests against the hemolyzed solution. To 4.5 ml. of a 10 per cent suspension of fractions. In all instances the indirect Coombs test was washed RBC was added 0.5 ml. of the diluted trypsin negative. solution. This mixture was incubated for 10 minutes at The influence of hemolysis per se on the Coombs test 370 C., centrifuged at 2500 rpm. for 10 minutes and the was appraised also by the injection of 300 cc. of hemo- supernatant decanted. The cells were re-suspended to 2 lyzed blood, obtained by repeated freezing and thawing, per cent suspension in the buffered saline before using. into a previously unused normal dog. The effect of this The test for "incomplete antibodies" using a colloid procedure on the direct Coombs test, Heinz body con- medium (34) was performed by the use of a final sus- centration and mechanical fragility of the recipient's pension of erythrocytes in 22 per cent human albumin circulating erythrocytes was noted periodically for one solution according to the following steps: The 22 per week. These tests yielded negative results. cent solution was made up from a 25 per cent solution These observations indicated that hemolysis per se of human albumin in buffered diluent (E. R. Squibb and did not cause a false positive Coombs test with the ca- Sons, New York) by adding buffered saline (pH 7.3 to nine antiserum prepared in this laboratory. 7.5). To 0.1 cc. of 22 per cent albumin in a 13 X 75 mm. test tube was added 1 drop of a 2 per cent suspension of RESULTS the test cells which had been washed three times in abun- Direct Coombs test dant saline. This was incubated at 37° C. for 30 min- utes, centrifuged at 500 to 1000 rpm. for 1 minute and ex- a) Results znith the undiluted Coombs serum amined macroscopically for clumping. If no gross ag- (Table I). The results from seven dogs fol- glutination was evident the contents of the tube were lowed up to 41 days after the injection of phenyl- poured on a slide and examined microscopically. The eluates from the Coombs positive erythrocytes hydrazine are summarized in Table I. The direct were prepared by an adaption of the Landsteiner-Miller Coombs test became positive within two to three procedure (35) as follows: 15 to 20 cc. of defibrinated days. The gross agglutination varied between 1 blood were centrifuged, the serum was removed, the and 3 +. The test was observed in these dogs to erythrocytes were washed four times with abundant sa- remain for 16 line, a final 50 per cent suspension of erythrocytes in positive days. It became negative saline was heated to- 560 C. with constant agitation for between 16 and 28 days. Dogs Nos. 6 and 7 6 to 8 minutes. The saline was then separated while (Table I) received additional doses of phenylhy- warm. As a result of this procedure hemolysis was mild drazine after the positive direct Coombs test had to moderate. The ability of such eluates to produce di- disappeared. The cycle of positive direct Coombs rect agglutination and agglutination by means of the test repeated itself in a manner similar ob- indirect Coombs procedure was tested by means of a to that panel of erythrocytes obtained from normal dogs. No served following the first injection. direct agglutination was noted. b) Results with serial dilutions of Coombs se- Since phenylhydrazine produces intravascular hemoly- rum with saline; the titer of Coombs serum; the sis and hemoglobinemia control tests were conducted in prozone phenomenon (Tables II, III, and IV). an attempt to determine whether hemolyzed canine blood Serial dilutions of per se could cause the erythrocytes obtained from normal the canine-serum anti-serum dogs to agglutinate under the influence of the canine anti- (Coombs serum from the rabbit) were prepared serum produced in the rabbit (canine Coombs serum) with saline. The diluted serum was added to and used in the experiments here reported. The normal washed 2 per cent suspensions of the canine eryth- dogs used in these tests were not used for the injection rocytes after the injection of phenylhydrazine. of phenylhydrazine. The procedure was For this purpose the indirect Coombs test was con- conducted periodically with ducted on the serum and hemolyzed erythrocytes obtained the red cells from dogs Nos. 1, 2, 3, 7, 8, 9, 10, and from two normal dogs. The serum and hemolyzed 11 as shown in Tables II, III, and IV. The dilu- erythrocytes were incubated with the erythrocytes ob- tions of 1: 8 to 1: 512 yielded positive results at 3 tained from three additional normal dogs and the eryth- to 6 days. Thereafter the titer gradually receded rocytes from the dog contributing the serum and hemo- lyzed erythrocytes. The incubation and the remainder except for dog No. 3, which expired and revealed of the indirect Coombs test was conducted as described a terminal elevation of the titer. POSITIVE DIRECT COOMBS TEST INDUCED BY PHENYLHYDRAZINE 1703

Absent or weaker agglutinations at lower dilu- serum were used. In the other four examples the tions followed by stronger agglutinations at higher phenomenon was prozone-like in that the change dilutions were noted in 8 of 32 titrations on the 8 was sudden from no agglutination to a positive dogs tested. This phenomenon had the character- result with higher dilutions. This blocking out istics of a prozone (36) on four occasions in the Of the positive test with the undiluted serum pos- sense of a gradual elevation in the intensity of ag- sibly explains the negative results noted with dogs glutination as succeeding higher dilutions of the Nos. 3, 5, and 7 on days 6, 13, and 5, respectively,

TABLE I Phenylhydrazine (40 mg. per kg.) was given following the observations on zero day *t

Direct Hema- Heinz Mech. Reticu- Dog No. Coombs Hb. tocrit body conc. fragility locytes Wt. kg. Day test gm./100 cc. % % % Hem. % 0 Neg. 11.0 36.4 0 4.4 0.3 1 Neg. 9.4 31.3 58.0 29.5 0.7 3 1 + 6.1 20.7 82.2 31.1 1.8 6 1 + 2.2 14.0 54.0 25.3 11.4 10 1 + 6.0 24.5 7.6 28.2 7.6 (1) 14 1 + 7.3 29.8 0.2 56.7 5.7 16 1 + 7.1 27.5 0.1 43.9 1.4 11 20 4 8.7 29.7 0 45.0 0.7 24 4 7.6 31.4 0 15.2 0.6 28 Neg. 8.9 29.4 0 67.3 30 i 10.8 34.5 0.1 33.9 0.1 35 Neg. 9.2 31.2 0 57.4 0.3 41 Neg. 11.8 32.4 0 17.0 0.5 0 Neg. 14.0 43.3 0 7.0 0.6 1 4 10.2 33.3 60.0 13.0 1.8 3 1 + 7.9 24.6 75.7 47.7 4.4 6 1 + 6.6 26.1 41.5 28.7 9.9 (2) 10 i 9.4 30.9 30.0 54.2 2.6 14 =1 9.8 37.2 1.5 58.3 2.9 6.4 17 4 10.6 36.3 0 35.2 1.1 20 4 12.0 39.2 0 6.2 0.5 24 Neg. 10.6 37.2 0 40.6 0.3 28 Neg. 10.8 39.1 0 81.6 30 i 12.5 40.4 0.1 24.8 0.1 35 Neg. 12.8 37.2 0.9 9.2 0.1 41 Neg. 13.5 43.5 0 40.0 0.4 0 Neg. 14.4 44.9 0 7.3 0.1 1 Neg. 10.0 37.2 72.0 12.5 0.4 3 1 + 7.2 23.0 85.6 73.4 1.3 (3) 6 Neg. 4.7 19.7 48.0 50.0 14.6 10 1+ 5.2 20.6 16.1 97.0 3.4 14 i1 5.2 22.8 0.3 21.3 1.9 5.5 17 3+ 6.0 23.1 0.2 40.8 2.9 Expired 2 1 + 14.0 47.0 3 1 + 7.9 26.0 60.0 6 3 + 3.5 16.0 64.0 7 3+ 52.0 (4) 12 2+ 9.0 13 1 + 5.8 11 15 1 + 6.4 27.2 4.7 16 Neg. 1.6 19 Neg. 7.5 29.9 1.4 21 Neg. 8.0 28 Neg. 0.7

* The direct Coombs test was obtained by the use of the undiluted canine serum anti-serum (Coombs serum). t Hb: Hemoglobin concentration. Heinz body concentration: Concentration Heinz body % of 1,000 RBC. Mechanical fragility: Mechanical fragility RBC % hemoglobin liberated. 1 704 E. t. MUIRHEAD, M. GROVES, AND S. BRYAN TABLE I-Continued

Direct Hema- Heinz Mech. Reticu- Dog No. Coombs Hb. tocrit body conc. fragility locytes Wt. kg. Day test gm./100 cc. % % % Hem. % 2 1+ 11.5 A 27.5 54.0 3 1+ 8.4 23.0 61.0 6 2+ 5.1 20.0 47.0 8 3 + 39.0 (5) 12 1 + 30.0 5.0 13 Neg. 28.0 3.7 6.4 15 3 + 7.6 27.8 1.8 16 3+ 30.2 0.7 19 1 + 8.4 30.4 0.4 21 3+ 9.1 30.0 28 0.3 44t Neg. 13.1 41.0 0 5.9 45 i 11.6 40.9 74.0 22.0 48 4 8.0 30.2 77.4 42.6 51 i 11.2 34.3 31.8 11.8 (6) 56 4 10.6 33.9 6.7 37.8 9.0 66 1 + 11.0 34.0 0 27.0 69 i 10.6 34.1 0 76.4 85§ Neg. 10.6 34.1 0 5.6 87 i 7.1 36.4 83.6 33.8 91 1 + 5.8 21.8 57.9 45.1 93 1 + 7.6 25.5 23.6 10.3 98 i 8.4 30.4 2.7 55.5 0 Neg. 10.9 35.4 0 5.8 1 i 9.6 35.4 29.0 9.0 3 1 + 37.0 52.0 25.0 5 Neg. 9.6 32.0 32.0 14.3 7 2 + 7.1 26.3 16.0 3.2 10 1 + 7.1 29.7 1.5 6.5 13 1 + 10.2 35.6 0.2 8.2 (7) 18 Neg. 10.2 35.6 0 10.2 11.4 28 Neg. 11.8 38.3 0 5.4 47t Neg. 12.3 40.0 0 49 1 + 7.1 92.2 27.3 50 2 + 7.6 15.7 53 2 + 4.4 19.3 56.5 7.8 55 2 + 6.9 25.4 12.8 4.5 60 4t 8.1 30.4 6.7 13.8 t Second dose of phenylhydrazine 44 and 47 days after first dose. § Third dose phenylhydrazine 41 days after second dose. as shown in Table I. These negative results with Eventually, even with trypsin, the test became the undiluted serum were preceded and followed negative. by positive results. d) Neutralization of the Coombs serum from c) Potentiation of the positive direct Coombs the rabbit with canine serum (Table IV). The test zwith trypsin (Table III). Subjecting the Coombs serum was diluted serially with saline for erythrocytes to the action of trypsin prior to the one set of tests. For another set of tests on the addition of the canine serum anti-serum accentu- same erythrocytes the Coombs serum was diluted ated the degree of agglutination. The treated cells serially in the same manner with canine serum also carried the test in the titration of the serum prepared by pooling equal parts of serum obtained one to several tubes beyond that observed with- from four normal dogs. The Coombs serum and out the action of the enzyme as noted in eight normal canine serum were allowed to interact for comparisons conducted on dogs Nos. 1, 2, 7, and 30 minutes before use. The erythrocytes tested 9. Trypsin alone, under the conditions used, was yielded a positive direct Coombs test following the not able to affect agglutination of the erythrocytes. injection of phenylhydrazine. This procedure was POSITIVE DIRECT COOMBS TEST INDUCED BY PHENYLHYDRAZINE 1705

TABLE II Data on the Coombs titer obtained by serial dilution of the Coombs serum with saline and its addition to washed erythrocytes following the injection of phenyihydrazine

Serial dilutions of Coombs serum Dog No. Day 1:1 1:2 1:4 1:8 1:16 1:32 1:64 1:128 1:256 1:512 1:1024 1:2048 3 0 0 0 + + + + + 5 + + + + + 4 4 4 4 4 0 10 + + + + 4 4 4 0 0 0 (1) 14 + + + + 4 : i 4 0 0 0 \16 ++ 4 4- 4 4 0 0 0 20 4 + 4 4 i i 4 4 d °0 24 4 4 4 4 4 A 0 0 0 0 0 30 4 4t 0 0 0 0 0 0 0 0 0 6 + + + + + 0 0 0 + 4t 0 0 14 4 4:1: 0 0 0 0 0 0 0 (2) 17 4 4 4 0 0 0 0 0 0 0 20 4 0 0 0 0 0 0 0 0 0 24 0 0 0 0 0 0 0 0 0 0 0 3 + 0 0 0 0 0 0 0 + + ± 0 6 0 0 0 0 0 i 4 + at i 0 (3) 10 + 4 4 :1 4 + 0 0 0 14 i 4 4 + ++ + + + i 4- 0 17 +++ +++ +++ ++++++ ++ ++ ++ $ 0 0 0 Expired conducted a total of seven times on the erythro- conducted and minimized it on the seventh oc- cytes from four dogs (Dogs Nos. 8, 9, 10, and casion. This type of neutralization suggests that 11). The direct Coombs test was positive with antibodies in the Coombs serum reacted with cor- the saline dilutions of the Coombs serum of 1: 8 to responding canine proteins on the affected eryth- 1: 64. Dilution and incubation of the Coombs rocytes in causing the agglutination. serum with the pooled canine serum neutralized the e) Results with the serum from normal rabbits ability of the Coombs serum to cause agglutina- (Table IV). The washed erythrocytes which tion of the erythrocytes in six of the seven tests yielded a positive direct Coombs test with the

TABLE III Data on the Coombs titer obtained with untreated erythrocytes and with the same cells after treatment with trypsin

Serial dilutions of Coombs serum Coombs Dog Day method 1:1 1:2 1:4 1:8 1:16 1:32 1:64 1:128 1:256 1:512 1:1024 1:2048 24 Regular 4 4: 41 4_ 0 0 0 0 0 Trypsin ++ 4 4 4 4i 0 0 0 0 (1) 30 Regular 41 4- 1 0 0 0 0 0 0 0 0 0 Trypsin 4 4 :1: 0 0 0 0 0 0 0 0 24 Regular 0 0 0 0 0 0 0 0 0 0 0 Trypsin + :1 0 0 + + 0 0 0 0 0 30 Regular 4 0 0 0 0 0 0 0 0 0 0 0 Trypsin :41 4 0 0 0 0 0 0 0 0 0 2 Regular 0 1 4 4 A 4 0 0 0 0 0 (7) Trypsin + + + + A 4 4 0 0 0 6 Regular + + + + 4 4 4 0 0 0 0 0 Trypsin + + + + + + 4 4- 0 0 0 2 Regular + + + + + ++ + 4 4 4 0 Trypsin + + + + + + + + + + + 4 4 = 4t 8 Regular ++ ++ + + + + + + 4- 4- 0 0 0 0 Trypsin ++ ++ ++ ++ ++ ++ + + + 0 0 0 170)616E. E. MUIRHEAD, M. GROVES, AND S. BRYAN

TABLE IV The netralization of the Coombs se b meis of can serum is demonstred. Failure dired Combs positiwe erytaroes Jo agguit weta rabiW serum, normal canine serumof and concentra lbumin is also depicted

RBC RBC RBC Serial dilutions of Coombsem + + f DDog og Rab. Coc. No. Day Procedure 1:1 1:2 1:4 1:8 1:16 1:32 1:64 1:128 1 t256 1:512 1:1024 1:2048 serum ser. sib. 5 Coombsserum+ -1+ l 1+ 1+ + + + 0 0 0 0 0 0 0 0 Saline dilution Coombsserum+ 0 0 00 0 0 0 0 0 0 8 Dog serum dil. 7 Coombsserum+ k 0 0 0 0 0 0 0 0 0 Dog serum dil. 4 Coombsserum+ + 1+ 1+.1+ + 0 0 0 0 0 0 0 0 0 Saline'dilution

Coombsserum+ 2+ 4: 1 4 - 0 0 0 0 0 Dog serum dil. 3 Coombsserum+ )1+ 1-I1+'~;00+0 0 0 0 0 0 Saline dilution Coombs serum+ 0 0 0 0 Dog serum dil. 10 4 Coombs serum + 4 4 4 0 0 0 0 0 0 0 0 0 0 0 0 Saline dilution Coombs serum + 0 0 0 0 0 0 0 0 0 0 0 0 Dog serut dil. 3 Coornbsserum+ d 1+ 2+ 1+ 2+ 1+ 4 0 0 0 0 0 Saline dilution Coomb serum+. 0 Dog serum dil. 4 Coombsserum+ 1+ ,+ 2+ 2+ 1+ 1+ A 0 0 0 0 O O O Saline dilution Coombsserum+ 0 .0 0 0 0' 0 0 0 0 Dog serum dil. saline dilutions of the Coombs serum up to 1: -8 erythrocytes also failed to agglutinate in the serum to 1 : 64 failed to agglutinate when placed in the from a normal dog and in a concentrated human serum from normal rabbits. The same washed albumin solution.

TABLE V The albumin test Results of the indirect Coombs test utilizing a panel of 3 or 4 It has been demonstrated that "incomplete anti- erythrocytes from normal dogs and the serum and also be detected etuate from the blood c bodies" coating erythrocytes can Coombs test was pP rominetmypositiec by agglutination within a colloid medium, such - as 22 per cent serum albumin. The albumin test Indirect Panel RBC following phenylhydrazine was consistently nega- Dog Direct Coombe No. DaV Coombs with 1 2 3 4 tive while the direct Coombs test was positive. (3) 17 3+ Serum 41 1 1+ This observation was made repeatedly with the S- ate 40 0 1+ erythrocytes-of eight dogs (Nos. 1, 2, 3, 7, 8, 9, (4) 7 3+ Eluate 0 0 0 ° 0, and 11). (5) 7 3+ Serum 0 0 0 Eluate 4 Elutioncof direct Coombs positive erythrocytes (9)* 10 2+ Serum 1+ ~1+ 4 (Tables V and VI) Eluate- 4 1 + 1 + Eluates were prepared by the Landsteiner- of six * Second dose. Miller procedure from the erythrocytes POSITIVE DIRECT COOMBS TEST INDUCED BY PHENYLHYDRAZINE11707

TABLE V The Coombs titer as obtained by serial dilution of the Coombs seram wih saline and the addition of this serum to the erythro- cytes of three normal dogs after they had been incubated with the serum and erythrocytic eluate both obtained at the time thathhe direct Coombs test was positive

Serum (S) Serial dilutions of Coombs serum with saline RBC or No. Eluate (E) 1:1 1:2 1:4 1:8 1:16 1:32 1:64 1:128 1:256 1:512 1:1024 1:2048 1 S + + + + . d 4 i 4 0 0 E 4 4 *+ + + + + + 1 A 1 0 2 S + + 4+ E -i1 :1 0 0 E + + + + + + -4 A 0 0

3 S - 4:* 4A 4 0 0 0 0 0 E + + + + 4 4 -4- 0 0 0 0 dogs (Nos. 1, 2, 3, 4, 5, and 9) at the time the di- against certain red cells from other animals; 2) rect Coombs test was positive with the undiluted that the material in the serum causing the positive Coombs serum. The eluate and the serum ob- indirect Coombs test was derived from the erythro- tained at the same time were tested by means of cytes altered by the drug and yielding a positive the indirect Coombs test against the erythrocytes direct Coombs test, just as a positive test was ob- obtained from three or four normal dogs. In tained by elution; and 3) that the drug altered Table V the indirect Coombs test was conducted normal proteins in the plasma, thus making these with undiluted Coombs serum while in Table VI adsorb onto erythrocytes. serial dilutions of Coombs serum with saline were used in the indirect Coombs test as applied to the Relation of general hematologic features following serum and eluate of one dog (No. 9). The eluates phenylhydrazine to the positive direct Coombs of two of the four dogs tested yielded positive in- test (Table I) direct Coombs test. The eluates and the cor- a) Relation to the anemia. It is apparent in responding sera gave similar results. These ob- Table I that the magnitude of the positive direct servations indicate that the causative agent of the positive direct Coombs test following phenyl- TABLE VII hydrazine is elutable but does not elucidate the Data on the indirect Coombs test utilizing the serum following phenylhydrazine and a panel of erythrocytes nature of the eluted material. obtained from three normal dogs

The indirect Coombs test (Table VII) Indirect Coombs test panel cells The indirect Coombs test was applied to the Dog Day 1 2 3 serum of the dogs receiving phenylhydrazine uti- O 0 0 + lizing the same panel of erythrocytes from three 4 4 . + 7 1 0 Hem. normal dogs throughout each experiment. It -11 + X + should be noted in Table VII that the indirect (1) 15 + + + 17 0 0 :4: Coombs test was negative in all but one case prior 21 + :4 + to the injection of the drug but that it subsequently 25 i1 1 41 became slightly positive in all cases. Although o 0 0 0 not 4 1 41 0 these observations have been elucidated, they 7 0 0 0 merit some comment. (2) 15 = .+ A 18 + 1: The development of the positive indirect Coombs 21 :1 4: + test to certain red cells cannot be ascribed to iso- 25 + 4 4 immunization since no blood transfusions were in- O 0 4 0 volved. Three alternative explanations for this 4 + + + (3) 7 0 0 0 phenomenon may be considered: 1) That the drug 11 + + ++ or its hemolytic effects potentiated or unmasked 15 + + + 18 41 4 + the presence of low titered natural antibodies 1708 E. E. MUIRHEAD, M. GROVES, AND S. BRYAN Coombs test bore no constant relationship to the in four of five dogs so studied, at which time the degree of anemia induced by the phenylhydrazine. direct Coombs test was positive. This degree of Thus the positive test not only occurred at height mechanical fragility then receded as the direct of the anemia, as demonstrated by the hemoglobin Coombs test tended to become negative, but in all concentration and the hematocrit reading, but per- dogs so studied (Nos. 1, 2, 3, 7, and 9) the fra- sisted and at times became more prominent dur- gility reached additional maxima and at times ing recovery from the anemia. (dogs Nos. 1, 2, 7, and 9) when the direct Coombs b) Relation to Heinz bodies. In these studies test became negative. the appearance and disappearance of the Heinz d) Relation to reticulocytes. The reticulo- bodies within the erythrocytes, as expressed by cyte count (dogs 1, 2, and 3) did not become the percentage of erythrocytes containing these prominently elevated until the anemia reached its structures, resembled that described by Cruz (37). maximum. There seemed to be a lag of 3 to 6 Thus the Heinz body concentration reached a days before the upsurge of reticulocytes, despite maximum in 3 to 6 days, and receded to the base the prominent stimulus of hemolysis. The reticu- line in 13 to 18 days. There was no constant re- locytosis receded at about 15 days. Thus the lationship between the proportion of erythrocytes reticulocytosis was prominent during the interval containing Heinz bodies and the reactivity of the when the direct Coombs test was positive; how- erythrocytes with Coombs serum. The direct ever, at times the Coombs test remained positive Coombs test remained positive after the concen- or increased in intensity (dog No. 3) after the tration of Heinz bodies receded to or near ex- reticulocytosis had receded. tinction. This point was emphasized further in e) Relation to hemoglobinemia and jaundice. two dogs in which the Heinz body concentration The positive direct Coombs test was questionable was determined at 1.5, 3, 4.5, and 24 hours fol- or positive at the height of the hemoglobinemia in- lowing the injection of the drug. The concentra- duced by phenylhydrazine but remained the same tion of Heinz bodies at 4.5 hours was 28 per cent or became more positive after the recession of the for both dogs and at 24 hours the concentrations elevated concentration of plasma hemoglobin. were 37 and 53 per cent. Yet during the entire Jaundice was not prominent in these experiments interval, the direct Coombs test remained nega- and as a consequence appeared unrelated to the tive, even when serial dilutions of the Coombs positive direct Coombs test. serum with saline were used to avoid the prozone phenomenon. Direct Coombs test following addition of phenyl- The Heinz body has been interpreted to re- hydrazine to blood in the test tube sult from alteration in the structure of the red Phenylhydrazine produces hemolysis and Heinz cell (38, 39). According to Ponder (39) the bodies within the erythrocytes when the drug is structure consists mainly of denatured globin. added to blood in the test tube, thus demon- Since it appears to possess a protein component strating that the drug injures red cells in vitro. the possibility that the altered protein structure This finding suggested the need to test the direct might attract the canine serum anti-serum of the Coombs test following action of the drug in the rabbit required consideration. These observations test tube. Accordingly, to 1 cc. aliquots of oxa- indicate that the Heinz bodies per se were not the lated canine blood was added 0.5, 0.25, 0.1, and structures causing the positive direct Coombs 0.05 mg. of phenylhydrazine. The first three test. tubes were allowed to stand at room temperature c) Relation to mechanical fragility of erythro- for 0.25, 1.0, 2.5 to 3.0 hours, respectively, and cytes. Following the injection of phenylhydrazine the last tube was kept in a 370 C. water bath for the mechanical fragility of fresh erythrocytes was 18 to 24 hours. At the end of these intervals the prominently elevated (Table I) from an average erythrocytes were checked for the presence of control value of 5.8 per cent hemolysis (4.4 to 7.0 Heinz bodies and were washed and subjected to per cent) to values varying between 25 and 97 the direct Coombs test. per cent, usually 57 to 77 per cent. The mechani- Blood obtained from six normal dogs was so cal fragility reached a maximum at 11 to 15 days treated. All preparations revealed the presence POSITIVE DIRECT COOMBS TEST INDUCED BY PHENYLHYDRAZINE 1709 of Heinz bodies. In the gross appraisal of the Coombs test in the dog by infusing incompatible direct Coombs test the erythrocytes in 10 of the 23 blood and in canine hemolytic disease of the new- tests appeared slightly to moderately clumped born in which "incomplete antibodies" were im- (1 + to 2 + results) but upon microscopic in- plicated as the causative factor. Chute and Som- spection the clumps disintegrated after 1 to 3 mers (46) observed a positive indirect Coombs minutes. The aggregates, which corroborated the test in parabiotic disease of the rat. The finding gross appearance, appeared to be true clumps and was interpreted as resulting from erythrocytic did not resemble coarse rouleaux or so-called red incompatibility causing "incomplete antibodies" cell "drifts" (4). This transient agglutination in one parabiont which were exchanged through suggests a tendency following treatment in the the common circulation and affected the erythro- test tube toward the same phenomenon observed cytes of the partner. after the intravenous injection of the drug. Fur- Three important features of the pattern of de- ther observations are required to clarify this pos- velopment of the positive Coombs test in these sibility. A distinct difference, however, was noted experiments militate strongly against an antigen- between the agglutination in the direct Coombs antibody interpretation, and emphasize the weak- test following the injection of phenylhydrazine ness of this hypothesis as a general explanation for into dogs, for in these instances the agglutination all instances of positive Coombs test. First, the remained stable for the period observed, which rapid appearance of the positive Coombs test, always exceeded 5 minutes and in some special within 2 days following injection of the drug, con- instances exceeded one hour. stitutes a strong objection to an antigen-antibody response. Second, the short-lived state of the posi- DISCUSSION tive test, as indicated by its disappearance in two It is commonly assumed that erythrocytes yield- to four weeks, is such as to be unusual for an anti- ing a positive direct Coombs test are coated with body response. Finally, the lack of potentiation antibodies of the "incomplete type" (1-20). This of the test by subsequent doses of the drug and interpretation, dependent on an antibody response the seemingly stereotyped response to subsequent to an antigen, appears definite in certain condi- doses do not connote immunologic characteristics. tions, such as hemolytic disease of the newborn The drug used, phenylhydrazine, is known to (4-6) and following certain incompatible blood produce direct injury to erythrocytes which is as- transfusions (7). In other states the antibody sociated not only with hemolysis but also with al- interpretation for a positive direct Coombs test terations within the erythrocytes, as demonstrated is attractive although not as well founded. Thus, by the production of methemoglobin and Heinz in the acquired hemolytic anemias the positive di- bodies. The pattern of observations suggests that rect Coombs test has been considered to be a phenylhydrazine alters the red blood cells in such consequence of autoimmune mechanisms (8-17, manner as to change their capacity to react with 40). This concept has gained considerable sup- the canine serum anti-serum of the rabbit. Three port by the demonstration of specificities of cer- theoretical possibilities for these alterations re- tain of these "antibodies" in acquired hemolytic quire further experimentation. In one the drug anemia (41, 42). Attractive as this hypothesis can be considered to alter the surface of the is the possibility of the production of abnormal erythrocyte so as to make it act directly as a re- globulins capable of attachment to erythrocytes ceptor for the canine antibodies. In another the by non-antigenic pathways cannot be ignored (21, drug can be considered to alter the surface of the 22). erythrocyte so as to cause it to adsorb non-im- While the literature on the appraisal of the di- mune proteins, as normal plasma proteins. Fi- rect Coombs test in man is extensive, few obser- nally, it is conceivable that the drug itself acts vations on its production in experimental animals as the coupling agent between the surface of the have been cited in support of the antibody hy- erythrocyte and non-immune proteins, as plasma pothesis. Noteworthy in this respect are the ob- proteins. servations of Young and associates (43-45) on While certain features of this study may be the production of a positive direct and indirect considered to constitute objections to an anti- 1710 E. E. MUIRHEAD, M. GROVES, AND S. BRYAN body response as the cause for the positive direct 2. The phenomenon does not appear to re- Coombs test, the findings of positive indirect sult from immunologic mechanisms for the rea- Coombs test by the use of the serum and eluates sons discussed. from the erythrocytes have been cited in other 3. The positive direct Coombs test appears to circumstances as supporting an immune response. result from an alteration of the erythrocyte by It is difficult to accept such an explanation from phenylhydrazine or some breakdown product of the present results, inasmuch as the pattern of phenylhydrazine. development of the positive Coombs test in these experiments has, for reasons previously cited, features which cannot be ascribed to antigen-anti- ACKNOWLEDGMENT body reactions. An alternative hypothesis would The authors are indebted to Drs. L. E. Young and S. N. Swisher for a batch of canine antiglobulin serum ascribe the positive indirect Coombs test to the prepared in their laboratory and used to test the activity same general process of action by the drug which of the canine anti-serum prepared for these experiments. seems responsible for the positive direct Coombs We are also grateful to Lt. Col. Joe Ackeroyd and Dr. test. Thus, according to this view, the factors D. W. Seldin for helpful suggestions in the preparation causing the positive direct Coombs test might be- of this material. come detached from the erythrocytes, pass into REFERENCES the serum and be reabsorbed by other erythrocytes. 1. Coombs, R. R. A., Mourant, A. E., and Race, R. R., It is not clear, as yet, what these factors are. A new test for the detection of weak and "incom- They may be degradation products of injured plete" Rh agglutinins. Brit. J. Exper. Path., 1945, erythrocytes or altered normal proteins. 26, 255. 2. Race, R. R., and Sanger, R., Blood Groups in Man. These considerations, based on the observations Springfield, Ill., Charles C Thomas, 1950. available in the literature and the observations 3. Muirhead, E. E., Problems in blood transfusion. herein recorded, make possible a reappraisal of III. Minor incompatible blood transfusions and the mechanisms evoking a positive direct Coombs acquired hemolytic anemia. Texas State J. Med., test. It seems definite that a positive direct 1954, 50, 465. 4. Coombs, R. R. A., Mourant, A. E., and Race, R. R., Coombs test may result from established antigen- In-vivo isosensitisation of red cells in babies with antibody reactions, as in hemolytic disease of the haemolytic disease. Lancet, 1946, 1, 264. newborn and following certain incompatible blood 5. Hill, J. M., and Haberman, S., Demonstration of transfusions. In the acquired hemolytic anemias, Rh antibodies in the newborn and further evi- idiopathic and certain secondary types, the af- dence of the pathogenesis of erythroblastosis. J. Lab. & Clin. Med., 1946, 31, 1053. fected individual's erythrocytes are coated with 6. Levine, P., Recent developments in iso-immuniza- globulins which resemble "incomplete antibodies." tion by the Rh factor. Am. J. Obst. & Gynec., This has led to the interpretation that the globu- 1945, 49, 810. lins are autoimmune, although the presence of ab- 7. Mollison, P. L., Blood Transfusion in Clinical Medi- normal non-immune globulins capable of adsorp- cine. Springfield, Ill., Charles C Thomas, 1951, pp. 240, 286, 222. tion to erythrocytes has not been excluded. The 8. Boorman, K E., Dodd, B. E., and Loutit, J. F., present experiments indicate the production of Haemolytic icterus (acholuric jaundice) congeni- the positive direct Coombs test by the action of a tal and acquired. Lancet, 1946, 1, 812. as phenylhydrazine. The findings suggest 9. Sturgeon, P., A new antibody in serum of patients drug, with acquired hemolytic anemia. Science, 1947, an action of the drug on the erythrocytes as the 106, 293. cause for this phenomenon. 10. Evans, R. S., Duane, R. T., and Behrendt, V., Dem- onstration of antibodies in acquired hemolytic ane- SUMMARY AND CONCLUSIONS mia with anti-human globulin serum. Proc. Soc. Exper. Biol. & Med., 1947, 64, 372. 1. The drug phenylhydrazine has produced a 11. Evans, R. S., and Duane, R. T., Acquired hemolytic positive direct Coombs test. The agglutination anemia. I. The relation of erythrocyte antibody has been stable when the drug was injected intra- production to activity of the disease. II. The sig- nificance of thrombocytopenia and leucopenia. venously into dogs and transient and unstable Blood, 1949, 4, 1196. when the drug acted on the blood in the test tube. 12. Dacie, J. V., and de Gruchy, G. C., Auto-antibodies POSITIVE DIRECT COOMBS TEST INDUCED BY PHENYLHYDRAZINE 1711

in acquired hamolytic anemia. J. Clin. Path., of hemoglobin in minute amounts of blood by Wu's 1951, 4, 253. method. J. Biol. Chem., 1931, 92, 589. 13. Dameshek, W., Acquired hemolytic anemia, physio- 30. Malloy, H. T., and Evelyn, K. A., The determination pathology with particular reference to autoim- of bilirubin with the photoelectric colorimeter. munization and therapy, in Proceedings of the J. Biol. Chem., 1937, 119, 481. Third Internatidnal Congress of the International 31. Mourant, A. E., see reference No. 2, p. 171-172. Society of Hematology, Cambridge, Eng., 1950, 32. Technical methods and procedures of the American New York, Grune and Stratton, Inc., 1951, 120. Association of Blood Banks, Minneapolis, Minn., 14. Wiener, A. S., Samwick, A. A., Morrison, M., and Burgess Publishing Co., 1953, p. 34. Loewe, L., Acquired hemolytic anemia. Am. J. 33. Morton, J. A., and Pickle, M. M., Use of trypsin in Clin. Path., 1952, 22, 301. the detection of incomplete anti-Rh antibodies. 15. Rosenfield, R. E., Vogel, P., and Rosenthal, N., The Nature, 1947, 159, 779. antiglobulin test; technic and practical applications. 34. Diamond, L. K., and Denton, R. L., Rh agglutination Am. J. Clin. Path., 1951, 21, 301. in various media with particular reference to the 16. Davidsohn, I., and Oyamada, A., Specificity of auto- value of albumin. J. Lab. & Clin. Med., 1945, 30, antibodies in hemolytic anemia. Am. J. Clin. Path., 821. 1953, 23, 101. 35. Landsteiner, K., and Miller, C. P., Jr., Serological 17. Dacie, J. V., Acquired hemolytic anemia, with special studies on the blood of the primates. II. The reference to the antiglobulin (Coombs') reaction. blood groups in anthropoid apes. J. Exper. Med., Blood, 1953, 8, 813. 1925, 42, 853. 18. Kidd, P., Elution of an incomplete type of antibody 36. Van Loghem, J. J., Kresner, M., Coombs, R. R. A., from the erythrocytes in acquired haemolytic and Roberts, G. F., Observations on a pro- ansemia. J. Clin. Path., 1949, 2, 103. zone phenomenon encountered in using the anti- 19. Komninos, Z. D., and Rosenthal, M. C., Studies on globulin sensitisation. Lancet, 1950, 2, 729. antibodies eluted from the red cells in auto- 37. Cruz, W. O., Acetylphenylhydrazine anemia. I. The immune hemolytic anemia. J. Lab. & Clin. Med., mechanism of erythrocyte destruction and re- 1953, 41, 887. generation. Am. J. M. Sc., 1941, 202, 781. 20. Young, L. E., and Miller, G., The long-term picture 38. Webster, S. H., Heinz body phenomenon in erythro- in autoimmune hemolytic disease. Tr. A. Am. cytes. A review. Blood, 1949, 4, 479. Physicians, 1953, 66, 190. 39. Ponder, E., Hemolysis and Related Phenomena. New 21. Wagley, P. F., Shen, S. C., Gardner, F. H., and York, Grune and Stratton, 1948, p. 363. Castle, W. B., Studies on the destruction of red 40. Young, L. E., Miller, G., and Christian, R. M., Clini- blood cells. VI. The spleen as a source of a sub- cal and laboratory observations on autoimmune stance causing agglutination of the red blood cells hemolytic disease. Ann. Int. Med., 1951, 35, 507. of certain patients with acquired hemolytic jaun- 41. Weiner, W., Battey, D. A., Cleghorn, T. E., Marson, dice by an antihuman serum rabbit serum (Coombs' F. G. W., and Meynell, M. J., Serological findings serum). J. Lab. & Clin. Med., 1948, 33, 1197. in a case of haemolytic anemia, with some general 22. Castle, W. B., The blood and blood-forming organs observations on the pathogenesis of this syndrome. in Brit. Med. J., 1953, 2, 125. Year Book of Medicine, Chicago, The Year 42. Book Publishers, 1950, p. 360; 1951, p. 277. Dacie, J. V., and Cutbush, M., Specificity of auto- antibodies in acquired hamolytic anaemia. J. Clin. 23. Clegg, J. W., and King, E. J., Estimation of hamo- Path., 1954, 7, 18. globin by the alkaline haematin method. Brit. 43. Med. Christian, R. M., Ervin, D. M., and Young, L. E., J., 1942, 2, 329. Observations on the in-vitro behavior of dog iso- 24. Wintrobe, M. M., A simple and accurate hematocrit. antibodies. J. Immunol., 1951, 66, 37. J. Lab. & Clin. Med., 1929, 15, 287. 44. Christian, R. M., Stewart, W. B., Yuile, C. L., Ervin, 25. Chaplin, H., Jr., and Mollison, P. L., Correction for D. M., and Young, L. E., Limitation of hemolysis plasma trapped in the red cell column of the he- in experimental transfusion reactions related to matocrit. Blood, 1952, 7, 1227. depletion of complement and isoantibody in the 26. Wintrobe, M. M., Clinical Hematology. 3rd ed., recipient Observations on dogs given successive Philadelphia, Lea & Febiger, 1951, p. 83. transfusions of incompatible red cells tagged with 27. Webster, S. H., Liljegren, E. J., and Zimmer, D. J., radioactive iron. Blood, 1951, 6, 142. Rapid staining of Heinz bodies in smears. Stain 45. Young, L. E., Christian, R. M., Ervin, D. M., Davis, Technol., 1948, 23, 97. R. W., O'Brien, W. A., Swisher, S. N., and 28. Shen, S. C., Castle, W. B., and Fleming, E. M., Ex- Yuile, C. L., Hemolytic disease in newborn dogs. perimental and clinical observations on increased Blood, 1951, 6, 291. mechanical fragility of erythrocytes. Science, 1944, 46. Chute, R. N., and Sommers, S. C., Hemolytic dis- 100, 387. ease and polycythemia in parabiosis intoxication. 29. Bing, F. C., and Baker, R. W., The determination Blood, 1952, 7, 1005.